Speech recognition systems and other natural language processing systems are being developed to enable increasing use of speech based interaction with mobile communication devices such as, for example, smartphones and vehicle phones. The environment in which these types of phones are used presents challenges to the speech recognition systems due to audio interference including wind noise, road noise and car radio, audio player or other in-vehicle entertainment systems. The phone's microphone typically picks up these interfering audio sources, along with the user's speech, thereby decreasing the reliability of the speech recognition system or other applications that may be employed to process or act on the user's speech. Additionally, the interfering audio generally makes it more difficult for the person at the other end of the phone call to understand the user's speech and follow the conversation.
Some systems attempt to cancel out the interference using processing techniques that require a-priori information, for example, probability and statistical information about the interfering signal. This may be possible to some extent for cancelling wind or road noise, which can be modeled as white noise with assumed statistical properties. Interfering audio signals, however, which may also include speech, are generally not amenable to these cancellation techniques.
Other systems attempt to address the interference problem by muting the vehicles speakers while the phone is in use, but this adversely impacts other passengers in the car by depriving them of the audio or entertainment that would otherwise be available.
Although the following Detailed Description will proceed reference being made to illustrative embodiments, many alternatives, modifications, and variations thereof will be apparent to those skilled in the art.